Porosity reduction during mechanical compaction of a sediment generally has been assumed to be controlled by the increase in vertical effective stress, which is convenient because vertical stress profiles may be readily calculated from density logs. Poroelasticity theory shows, however, that mean effective stress controls porosity reduction. According to published data, horizontal stresses increase with overpressure, as well as with depth, so mean stress and vertical stress profiles are poorly correlated in overpressured sections. We have used wireline logs to compare the pore pressures estimated in mudrocks by relating porosity to mean effective stress and to vertical effective stress for overpressured Tertiary sections in southeast Asia. Wells from three different basins were studied. Mudrock porosities were estimated from the sonic log response and sorted by lithology according to the natural gamma-log response. Two sets of normal compaction curves, relating porosity to mean effective stress and to vertical effective stress, were determined empirically by fitting data points where the pore pressure was thought to be hydrostatic. These curves were then used to estimate the minimum pore pressure corresponding to mudrock porosity values in the overpressured sections. The pore pressures inferred using the mean effective stress are consistent with direct measurements of pore pressure in the adjacent sands. In contrast, pore pressures inferred in mudrocks using the vertical effective stress are significantly lower for the overpressured sections, implying discontinuities in the pore pressure profiles at lithological boundaries, which cannot readily be explained. We conclude that the pore pressures estimated using the vertical effective stress are wrong and that empirical relationships between porosity and vertical effective stress should not be used for estimating pore pressures: porosity should be empirically related to mean effective stress instead.